[antlr-interest] question about antlr

[Bart Kiers]

Hi Patrick,

Sorry, I didn't express myself clearly. What I meant was if you could post a
stripped down version of your grammar that includes the problematic rule(s)
and also include a little test class that parses the example input. It
really helps if you post the exact input that produces the error, not
"something like this".
An example in the line of an SSCCE (http://sscce.org/) would be ideal.
I could do this all myself: creating a test class, adjusting the example
input and letting my test class parse it, but that'd take up quite a bit of
my time.

Regards,

Bart Kiers.


On Fri, May 27, 2011 at 9:42 AM, Patrick Hofman <
patrick.hofman at invantive.com> wrote:

>  Hi Bart,
>
>
>
> The input is something like this: '$C{D,.,.,.+1} + $C{D,.,.,.+2}'
>
>
>
> The problem is filtering out the ' + ' part of the expression. The
> expression goes well when I don't include the ' + ' part. In fact, this part
> of the expression should be copied 'as is' to the output in the end.
>
>
>
> The grammar part that is the problem is this part I think:
>
>
>
> eca_kolom_expressie
>  : {input.LT(1).getText().equals("$"+"C")}? ECA_ID ECA_LACCOLADE
> draaimethode separator sheet separator column separator row ECA_RACCOLADE
> {System.out.println("Gaaf, draaimethode " + draaimethode); mode="EXCEL";
> System.out.println("Mode = " + mode);}
>  ;
> excel_fragment
>         : {mode.equals("EXCEL")}? *EXCEL_FRAGMENT* {mode="ECA";
> System.out.println("Mode = " + mode);}
>  ;
> excel_formula
>  : excel_fragment (eca_kolom_expressie excel_fragment)*
>  ;
>
> Where EXCEL_FRAGMENT is:
>
>
>
> EXCEL_FRAGMENT
>  : (options {greedy=false;} : .)
>  ;
>
> The determination of excel_formula (entry point) (specific: the
> EXCEL_FRAGMENT part) goes well when I put in TILDE at the start and end of
> EXCEL_FRAGMENT. So it looks like the greedy is too greedy. It eats up a part
> of eca_kolom_expressie.
>
>
>
> Is this better for you to understand? I am a ANTLR rookie so I'm sorry when
> it is hard to get it right for you.
>
>
>
> Regards,
>
>
>
>
>
> Patrick Hofman
>
>  ------------------------------
> *From:* Bart Kiers [bkiers at gmail.com]
> *Sent:* Thursday, May 26, 2011 5:56 PM
> *To:* Patrick Hofman
> *Cc:* antlr-interest at antlr.org
> *Subject:* Re: [antlr-interest] question about antlr
>
>  Hi Patrick,
>
>  I have a hard time understanding what it is you're trying to do.
> Instead of springing the entire grammar on us (or me), could you extract
> only those rules that are causing the problem(s)? And could you include some
> example input you'd like to match/parse?
>
>  Regards,
>
>  Bart Kiers.
>
>
> On Thu, May 26, 2011 at 4:04 PM, Patrick Hofman <
> patrick.hofman at invantive.com> wrote:
>
>> And now the grammar
>>
>> From: Patrick Hofman
>> Sent: donderdag 26 mei 2011 15:35
>> To: antlr-interest at antlr.org
>> Subject: question about antlr
>>
>> Hi all,
>>
>> I have bought the ANTLR book in order to learn ANTLR better, but I still
>> don't get how to fix one thing. I hope you can help me with that.
>>
>> I have a grammar that is used to parse our custom formula format to an
>> Excel formula. So when filling an Excel worksheet the string entered is
>> something like '$C{D,.,.,.+1}' which means 'one cell right from the current
>> cell'. You will understand that eventually we will get something like
>> '$C{D,.,.,.+1} + $C{D,.,.,.+2}', so add up the value of the first cell right
>> and the second cell right (in Excel when we are at A1 this results in '=B1 +
>> B2'.
>>
>> The problem is this: In the grammar I cannot find a way to 'eat up' the
>> text between two 'eca_kolom_expressie' strings (the '$C{...}' part).
>>
>> I already tried 'EXCEL_FRAGMENT' in a hundred ways, but none of them
>> seemed to work. ('TILDE (options {greedy=false;} : .) TILDE' seems to work,
>> but when removing the TILDEs it stops working)
>>
>> I have included the grammar. Can you point me in the right direction?
>>
>> Regards,
>>
>> Patrick Hofman
>> Senior Consultant
>> Invantive B.V.
>>
>>
>> List: http://www.antlr.org/mailman/listinfo/antlr-interest
>> Unsubscribe:
>> http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>>
>>
>